Thermochronological Evidence of Early Orogenesis, Eastern Pyrenees, France
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Thermochronological Evidence of Early Orogenesis,
Eastern Pyrenees, France
Sébastien Ternois, Margaret Odlum, Mary Ford, Raphael Pik, Daniel Stockli,
Bouchaib Tibari, Arnaud Vacherat, Vincent Bernard
To cite this version:
Sébastien Ternois, Margaret Odlum, Mary Ford, Raphael Pik, Daniel Stockli, et al.. Thermochrono-
logical Evidence of Early Orogenesis, Eastern Pyrenees, France. Tectonics, American Geophysical
Union (AGU), 2019, 38 (4), pp.1308-1336. �10.1029/2018TC005254�. �hal-02394226�
HAL Id: hal-02394226
https://hal.univ-lorraine.fr/hal-02394226
Submitted on 13 Aug 2021
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est
archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
CopyrightRESEARCH ARTICLE Thermochronological Evidence of Early Orogenesis,
10.1029/2018TC005254
Eastern Pyrenees, France
Key Points:
• Zircon (U‐Th)/He data record a
Sébastien Ternois1 , Margaret Odlum2 , Mary Ford1 , Raphaël Pik1 , Daniel Stockli2 ,
cooling phase during early Bouchaïb Tibari1, Arnaud Vacherat1 , and Vincent Bernard1
convergence in the Pyrenean
1
low‐relief retrowedge Centre de Recherches Pétrographiques et Géochimiques, UMR 7358 CNRS‐Université de Lorraine, Vandoeuvre‐lès‐
(Agly‐Salvezines block) Nancy, France, 2Department of Geological Sciences, Jackson School of Geosciences, University of Texas, Austin, TX, USA
• Forward and inverse thermal history
modeling reveals limitations in
current radiation damage‐annealing
models for He diffusion in zircon
Abstract In collisional orogens, distinguishing the thermal signature of early orogenesis from the
• Crustal section sequential preceding rift or from subsequent thermal events is a major challenge. We present an integrated
restoration demonstrates that geological and low‐temperature thermochronology study of the Paleozoic Agly‐Salvezines crustal block in
tectonic processes alone can
the retrowedge of the eastern Pyrenees (France). The northern Pyrenees preserves one of the best geological
generate rapid cooling during early
orogenesis records of a rift‐to‐collision transition. The Agly‐Salvezines block represents the inverted distal European
margin of an Aptian–Cenomanian rift system. Seventeen samples were collected throughout the external
Supporting Information: orogenic massif and analyzed for low‐temperature thermochronology: zircon (U‐Th)/He dating documents
• Supporting Information S1
the cooling history of the massif during the initiation and early phase of Pyrenean convergence, while apatite
• Data Set S1
(U‐Th)/He dating completes the record of plate collision. Using inverse and forward modeling of new
low‐temperature thermochronology data, we show that the Pyrenean retrowedge records two clear phases of
Correspondence to: orogenic cooling, Late Campanian–Maastrichtian and Ypresian–Bartonian, which we relate to early
S. Ternois, inversion of the distal rifted margin and main collision, respectively. An earlier, late Aptian–Turonian
seb.ter@hotmail.fr
cooling history is detected, possibly related to rifting and/or postrift. No cooling is evidenced during the
Paleocene during which tectonic quiescence is recorded in the adjacent Aquitaine retroforeland basin. Using
Citation: our low‐temperature thermochronology data and geological constraints, we propose a crustal‐scale
Ternois, S., Odlum, M., Ford, M., Pik,
sequentially restored model for the tectonic and thermal transition from extension to peak orogenesis in the
R., Stockli, D., Tibari, B., et al. (2019).
Thermochronological evidence of early eastern Pyrenees, which suggests that both thrusting and underplating processes contributed to early
orogenesis, eastern Pyrenees, France. inversion of the Aptian–Cenomanian rift system.
Tectonics, 38, 1308–1336. https://doi.
org/10.1029/2018TC005254
1. Introduction
Received 26 JUL 2018
Accepted 24 JAN 2019 The Pyrenean orogen was generated from late Santonian–early Campanian to middle Miocene by N‐S con-
Accepted article online 2 FEB 2019
vergence of the Iberian and European plates (Choukroune, 1989; Muñoz, 1992; Macchiavelli et al., 2017).
Published online 10 APR 2019
External orogenic zones and foreland basins record two distinct periods of low but accelerating tectonic
shortening and subsidence, latest Santonian–Danian and Thanetian–Oligocene, separated by a quiet (very
low to near‐zero subsidence) period during the Paleocene (Ford et al., 2016). These two periods are
recognized to be synorogenic and mark two phases of convergence, the second generally recorded as
Eocene–Oligocene in low‐temperature thermochronology studies. Significantly, the orogen's three‐phase
convergence history does not correspond to behavior predicted by foreland dynamic models (Naylor &
Sinclair, 2008; Sinclair et al., 2005; Sinclair & Naylor, 2012). While the evolution of crustal thickening, ther-
micity, orogen relief, and erosion during the main Eocene–Oligocene collisional phase are well studied and
show a clear link to foreland basin dynamics (Fillon et al., 2013; Sinclair, 2011), the early Pyrenean history of
convergence and the subsequent phase of quiescence are poorly constrained. In particular, there are few data
on the thermal and dynamic behavior of the orogen during the earliest phase of Pyrenean convergence. The
record of early orogenesis is best observed in the retrowedge where shortening and translation values are low
(Naylor & Sinclair, 2008) and where the main Aptian–Cenomanian rifting phase is preserved. In this paper
we present new zircon and apatite (U‐Th‐Sm)/He (ZHe and AHe, respectively) data from an external
Paleozoic crustal block, Agly‐Salvezines (Figures 1 and 2), in the eastern Pyrenean retrowedge. By integrat-
ing these data with new and published geological data, we address the following questions: Can we recog-
nize a thermal signature for the onset of Pyrenean convergence and for Paleocene quiescence? Are early
crustal thermal events contemporaneous with early foreland tectonic events, or is there a lag time between
©2019. American Geophysical Union. these two phenomena? Can we distinguish an early orogenic thermal signature from other thermal events
All Rights Reserved. such as those associated with preceding rifting or later collision?
TERNOIS ET AL. 1308Tectonics 10.1029/2018TC005254
Figure 1. Simplified tectonic map of the Pyrenees showing the main tectonic units, the location of the study area (Figure 2), the position of the crustal‐scale cross‐
sectional model proposed in this paper (Figure 6), and the location of (i) the North Pyrenean basement massifs and (ii) the main plutons/gneiss domes in the
Axial Zone. NPFT = North Pyrenean Frontal Thrust; NPZ = North Pyrenean Zone; NPF = North Pyrenean Fault; SPFT = South Pyrenean Fault; MN = Montagne
Noire; MC = Massif Central; Mou = Mouthoumet; Co = Corbières. (i) L‐U = Labourd‐Ursuya; CV = Cinco Villas; Al = Aldudes; A‐M = Arbailles‐Mendibelza;
Ba = Barousse; Mh = Milhas; Ca = Castillon; Ar = Arize; TS = Trois Seigneurs; SB = Saint‐Barthélémy; Ag = Agly; Sa = Salvezines. (ii) A = Aya; EC = Eaux
Chaudes; CP = Cauterets‐Panticosa; Ne = Néouvielle; Bi = Bielsa; Bo = Bordères; Mil = Millares; Po = Posets; Ly = Lys‐Caillaouas; Ma = Maladeta;
Mm = Marimanha; Bas = Bassiès; As = Aston; Ho = Hospitalet; MtL = Mont‐Louis; Qu = Quérigut; Mi = Millas; Can = Canigou‐Carança; Sl‐Jo = Saint Laurent‐La
Jonquera.
The Pyrenean case study has both advantages and challenges for the study of early orogenesis. Advantages
include a low total convergence (92–165 km; Beaumont et al., 2000; Mouthereau et al., 2014; Muñoz, 1992;
Roure et al., 1989; Teixell et al., 2018, 2016) allowing better preservation of early phases than in high‐
convergence orogens. There is also no thermal signature of subduction preceding collision, as is the case in
many orogens, because no true oceanic crust was formed before the onset of convergence (Clerc &
Lagabrielle, 2014; Jammes et al., 2009; Lagabrielle et al., 2010; Lagabrielle & Bodinier, 2008; Masini et al.,
2014; Tugend et al., 2014). On the other hand, the Aptian–Cenomanian rifting generated a major thermal
pulse that had not reequilibrated before the onset of convergence some 10 myrs later (Angrand et al.,
2018). This inherited thermal perturbation may have been present during the first 30–35 myrs of orogenesis
(Angrand et al., 2018; Vacherat et al., 2014), maintaining temperatures during early orogenesis above the sen-
sitivity limit of low‐temperature thermochronometers (40–300 °C; e.g., Carrapa, 2010; Peyton & Carrapa,
2013) and thus delaying any cooling record until the main Eocene collision (Vacherat et al., 2016).
However, previous limited low‐temperature thermochronology data (Yelland, 1991) indicate that the massifs
at the eastern end of the external Pyrenean orogenic system, Agly‐Salvezines, record a Campanian–
Maastrichtian cooling signal, synchronous with the first period of accelerating subsidence in the retrofore-
land (Ford et al., 2016). This signature is absent in more proximal crystalline massifs further west, which only
record Eocene exhumation (e.g., Vacherat et al., 2016), with the exception of the Labourd‐Ursuya Massif at
the extreme western end of the orogen (Figure 1; Yelland, 1991; Vacherat et al., 2014). As paleotemperature
data obtained by Raman spectroscopy on carbonaceous material (RSCM; Clerc et al., 2015; Chelalou et al.,
2016, and references therein) and radiometric data (Clerc et al., 2015, and references therein) for Mesozoic
sediments suggest higher postrift cooling rates from end of rifting to the onset of convergence in the eastern
Pyrenean retrowedge, we can expect paleotemperatures at the onset of orogenesis to be lower in this region. A
TERNOIS ET AL. 1309Tectonics 10.1029/2018TC005254
Figure 2. Geological map of the Agly‐Salvezines area, based on maps published by the BRGM (Bureau de Recherches Géologiques et Minières; Quillan: Bessière
et al., 1989; Crochet et al., 1989; Tuchan: Berger, Alabouvette, et al., 1997; Berger, Bessière, et al., 1997; Rivesaltes: Berger et al., 1993; Fonteilles et al., 1993), showing
the main structural elements and lithologies. The position of the crustal‐scale sequentially restored cross‐sectional model proposed in this paper (Figure 6) is
indicated. Samples analyzed in this study are grouped into igneous/northern compartment rocks (yellow‐red symbol colors) and metamorphic/southern com-
partment rocks (purple‐blue symbol colors). Corresponding model results are shown in Figures 4, 5, S3, S5, and S6. Sample names underlined have been dated by
both ZHe and AHe methods.
clear low‐temperature thermochronology signal of the onset and early phase of orogenesis is therefore most
likely to be preserved in the east, specifically in the Agly‐Salvezines block.
Campanian‐to‐Maastrichtian successions of the early Pyrenean foreland basins record sediment supply from
a main eastern source area (Bilotte, 1985; Monod, 2014; Plaziat, 1981), which has since been destroyed by the
Oligocene–Miocene opening of the Gulf of Lion (Séranne, 1999; Gunnell et al., 2008, 2009). The early
orogenic wedge supplied little or no sediment at this time and therefore was of low relief, probably largely
submarine (Ford et al., 2016). During the Eocene main collision, the growth of major orogen relief then
TERNOIS ET AL. 1310Tectonics 10.1029/2018TC005254
migrated from east to west (Whitchurch et al., 2011, and references therein). Mass balance models predict
that orogen relief will grow as long as topographic uplift (due to crustal thickening) is not balanced by ero-
sion (growth phase; Willett et al., 1993; Willett & Brandon, 2002). Once this balance is achieved, the orogen
passes to a steady state phase. Numerical models that use thermochronology data to constrain the thermal
evolution of a growing orogen (e.g., Batt et al., 2001; Batt & Braun, 1997, 1999) have always evoked uplift
and erosion to interpret cooling events. However, in the case of early orogenic wedge growth, erosion is
expected to be nonsignificant. This could result in a negligible perturbation in thermal gradients at closure
temperature depths, therefore questioning the potential of low‐temperature thermochronology to record
cooling related to the onset of orogenesis. However, as the recent work of Mesalles et al. (2014) has high-
lighted, crustal thickening by early underthrusting alone may lower the geotherm of distal rifted margins
due to the downward deflection of isotherms. This implies that low‐temperature thermochronology data
may indeed record the initial stage of convergence.
In this paper we present a detailed low‐temperature thermochronology record of early convergence in the
eastern Pyrenees, more precisely in the Agly‐Salvezines region where the inverted distal margin of the
Aptian–Cenomanian hyperextended rift system is preserved. We show that the easternmost massifs of
the northern Pyrenees, unlike their more westerly counterparts, record two distinct cooling events during
convergence, Campanian–Maastrichtian (ZHe data) and Ypresian–Bartonian (AHe data), separated by a
period of very little cooling in the Paleocene. We argue that the early convergence cooling was due to early
orogenic wedge thickening by underplating of the Iberian distal margin beneath the Agly‐Salvezines crustal
block and coeval northward thrusting of this block onto the colder, northern European crust. Exhumation of
the thickening crust (wedge thickening by basal accretion coupled with erosion) is evoked as the principal
cooling mechanism for the Eocene main convergence cooling event. Erosion‐related exhumation then pro-
gressively migrated south. We discuss the implications of these new results for models of orogenic growth in
the Pyrenees and for early convergence in general.
2. Geological Setting
The Pyrenean mountain belt is a relatively narrow (100‐km‐wide), double‐wedged orogenic system trending
N110° and extending over 1,500 km from the Cantabrian platform in northern Spain to the Mediterranean
Sea (Languedoc‐Provence) in southwestern France (Figure 1). The orogen is divided into five tectonostrati-
graphic zones. The external orogenic zone on the southern Iberian plate comprises the South Pyrenean fold
and thrust belt and the Ebro (foreland) Basin (Vergés et al., 1995, 1998). The South Pyrenean Frontal Thrust
(SPFT) separates the two systems. The Axial Zone in the core of the Pyrenean range consists of south‐verging
thrust sheets of Variscan metasediments and basement (Barnolas et al., 1996; Muñoz, 1992, and references
therein). The Aquitaine (foreland) Basin and the North Pyrenean Zone form the retroforeland system on the
northern European plate separated by the North Pyrenean Frontal Thrust (NPFT; Biteau et al., 2006;
Souquet et al., 1977).
The Pyrenean chain was built on a complex crustal template. Basement blocks of the North Pyrenean
Zone, namely, the North Pyrenean basement massifs (Choukroune, 1974), and the Axial Zone consist
of Ediacaran–Paleozoic metasedimentary and Ordovician‐to‐Early Permian plutonic rocks recording
the Variscan orogeny (e.g., Barnolas et al., 1996, and references therein). This orogeny was followed
by two phases of Mesozoic extension or transtension, the first occurring from Triassic to Late Jurassic in
an east–west zone linking the Tethys and Central Atlantic oceans (e.g., Curnelle et al., 1980). Rifting
from late Aptian to early‐to‐middle Cenomanian (114–97 Ma) then created deep marine depocenters
(Debroas, 1987, 1990), adjacent to the North Pyrenean basement massifs and associated with exhumation
of mantle rocks (Azambre & Ravier, 1978; Clerc et al., 2012; Lagabrielle et al., 2010; Lagabrielle &
Bodinier, 2008).
3. Thermal History of the Pyrenean Orogeny
3.1. Pyrenean Thermal History
Extensive low‐temperature thermochronology studies in the Pyrenees over the last 20 years have produced a
world‐class data set to constrain crustal thermal behavior during orogenesis (Bosch et al., 2016; Hart et al.,
2017; Labaume et al., 2016; Mouthereau et al., 2014, and references therein; Rushlow et al., 2013;
TERNOIS ET AL. 1311Tectonics 10.1029/2018TC005254
Thomson et al., 2017; Vacherat et al., 2014, 2016). Bedrock low‐temperature thermochronology (AHe
and apatite fission track [AFT]) data from the Axial Zone and the basement massifs in the central
North Pyrenean Zone record only the Eocene–Oligocene (50–20 Ma) main collision cooling history,
younging from east to west and from north to south (Mouthereau et al., 2014; Whitchurch et al., 2011).
In contrast, relatively little is known about the thermal history of the early phase of Pyrenean convergence
(latest Santonian to Danian). Sparse detrital AFT, ZHe and zircon fission track (ZFT) data from early,
synorogenic depocenters (southern Pyrenees: Filleaudeau et al., 2012; Fillon et al., 2013; Whitchurch
et al., 2011) indicate that significant relief already supplied southern depocenters. However, foreland basin
sedimentary records show that up to the Eocene, this relief lay to the east of the present‐day mountain range
(Bilotte, 1985; Monod, 2014; Plaziat, 1981). This suggests that an active orogenic zone existed further east for
which two early cooling events are probably recorded by limited detrital low‐temperature thermochronology
data (95–135 and 60–80 Ma; Filleaudeau et al., 2012; Whitchurch et al., 2011; 85–110 and 60–85 Ma;
Mouthereau et al., 2014; Vacherat et al., 2017). The early Pyrenean orogen itself (Central Pyrenees) probably
formed a low relief or even submarine edifice (Ford et al., 2016; Vacherat et al., 2017), with slow exhumation
(0.2 km/Myr) and little sediment yield until the late Eocene (Fillon & van der Beek, 2012; Fitzgerald et al.,
1999). We would therefore expect that any sediments sourced from this area would be restricted to the upper,
thermally non‐reset levels of the crust.
Nevertheless, an early (Campanian to Maastrichtian) cooling event is recorded in sparse bedrock ZFT and
ZHe data from Variscan basement massifs of the western (Labourd‐Ursuya; Vacherat et al., 2014; Yelland,
1991) and eastern (Agly; Yelland, 1991) North Pyrenean Zone. Several factors may contribute to the lack of
a clear record of early orogenic cooling. It is possible that sampling has been, as yet, too sparse to detect
this early event. However, early crustal thickening probably affected only the distal rifted margin, migrat-
ing outward to the proximal areas with time (definition of “distal” and “proximal” according to Sutra
et al., 2013). Another reason could be simply that the early record of convergence in the Central
Pyrenees was eroded away off the Axial Zone sometime from the Late Cretaceous to present (e.g.,
Filleaudeau et al., 2012; Rushlow et al., 2013; Thomson et al., 2017; Whitchurch et al., 2011). Only the
more distal areas would have preserved the record of early orogeny. In any case, basement of the distal
margin is poorly preserved in the orogen. Basement massifs in the eastern North Pyrenean Zone are, how-
ever, recognized as representing the distal upper plate (Agly and Salvezines; e.g., Clerc & Lagabrielle, 2014;
Clerc et al., 2016).
3.2. The North Pyrenean Zone
The North Pyrenean Zone is a 10‐ to 40‐km‐wide, north‐verging thrust system (Figure 1). This zone lies
between the NPFT to the north and the steeply dipping North Pyrenean Fault (NPF) to the south that is
generally accepted as the surface trace of the plate boundary (Choukroune, 1989). It consists of a
deformed Triassic to Turonian sedimentary succession, overlying Variscan basement that is exposed in
the North Pyrenean basement massifs (Baby et al., 1988; Barnolas et al., 1996; Choukroune, 1974, and
references therein).
The eastern North Pyrenean basement massifs show widespread hydrothermal alteration (250–550 °C) in
the form of albitization, dequartzification, and massive talc‐chlorite deposits dated middle Aptian to early
Turonian (117–92 Ma; Boutin et al., 2016, and references therein). Along its southern margin the North
Pyrenean Zone includes the Metamorphic Internal Zone, a narrow (0.5–10‐km‐wide) area of high‐
temperature (HT)‐low‐pressure (LP) metamorphism affecting Mesozoic sedimentary rocks (Choukroune,
1974; Ravier, 1957; Ravier & Thiébaut, 1982; Figure 1). The thermal perturbation is dated 113–85 Ma over
the whole Pyrenean realm, encompassing a major phase of extension and associated magmatic activity
(Golberg & Leyreloup, 1990; Montigny et al., 1986, and references therein). Centimeter‐scale fragments to
kilometer‐scale blocks of subcontinental mantle rocks and slices of granulitic crustal rocks occur throughout
the Metamorphic Internal Zone, often associated with, or encased in, carbonate breccias (Azambre & Ravier,
1978; Clerc et al., 2012; Lagabrielle et al., 2010; Lagabrielle & Bodinier, 2008). It is argued that this zone
represents an area of extreme crustal thinning, possibly hyperextension, where exhuming mantle provided
the heat source that caused metamorphism of the overlying Mesozoic strata (Chelalou et al., 2016; Clerc &
Lagabrielle, 2014; Clerc et al., 2015, 2016; Vauchez et al., 2013).
TERNOIS ET AL. 1312Tectonics 10.1029/2018TC005254
3.3. The Agly‐Salvezines Area
3.3.1. Structure and Stratigraphy
The Agly Massif is a N110° trending lens‐shaped half‐dome (35 km by 8 km), consisting in two main assem-
blages, hereafter referred to as the southern (mainly middle crust) and northern (middle‐upper crust) compart-
ments, separated by steep faults (Berger et al., 1993; Fonteilles et al., 1993; Figure 2). The Salvezines Massif
outcrops as a 3 km by 2 km tectonic window trending N110° and represents the western continuation of the
northern Agly compartment. Together, the two massifs are hereafter referred to as the Agly‐Salvezines block.
The Agly Massif preserves the most complete, albeit tectonically reduced (less than 10‐km‐thick), section of
Variscan crust outcropping in the Pyrenees (Barnolas et al., 1996, and references therein). The crustal section
is characterized by an overall apparent tilting NNE of its northern, upper compartment that may be Variscan
in age (Olivier et al., 2004, 2008). From south to north (or bottom to top) the massif consists of the following:
(i) the granulite facies Caramany Unit (paragneisses, marbles, orthogneisses, granitoids; Precambrian‐to‐
Lower Cambrian protoliths), typical of the base of the middle crust; (ii) the amphibolite facies Bélesta
Unit (partly migmatized paragneisses, orthogneisses, granitoids, and pegmatites; Precambrian‐to‐Lower
Cambrian protoliths), typical of the middle crust; and (iii) a metasedimentary cover of greenschist to amphi-
bolite facies rocks (Cambro‐Ordovician‐to‐Devonian protoliths), typical of the upper crust (Althoff et al.,
1994; Berger et al., 1993; Delay, 1989; Olivier et al., 2004, 2008; Tournaire Guille et al., 2018). Plutonic
intrusions include Lower‐to‐Middle Ordovician granitic laccoliths in the Paleozoic cover (Rivérole orthog-
neiss) and upper Carboniferous granitic bodies at the base of the Caramany Unit (Ansignan charnockite,
5–6 kbar, 20‐ to 23‐km depth; Andrieux, 1982; Vielzeuf, 1996) and in the Paleozoic cover (Saint‐Arnac
granite, 5‐ to 10‐km depth). Late Carboniferous plutonism is recognized to be closely linked to a well‐known,
albeit poorly dated, Carboniferous HT‐LP metamorphism (e.g., Barnolas et al., 1996). The Salvezines Massif
comprises from bottom to top the following: (i) upper Carboniferous granite and gneiss in its central and
southern part and (ii) Devonian–Visean metasedimentary rocks in its NNW sector. Salvezines gneisses
present facies and textural similarities to the Rivérole orthogneiss (Demange & Pascal, 1979; Pascal, 1979).
The Agly‐Salvezines block is surrounded by three N090°‐to‐N110° trending regional synforms of Keuper to
Apto‐Albian strata (Figure 2; Fonteilles et al., 1993). (i) The Boucheville synform to the SSW, immediately
north of the NPF, is separated from the oldest and highest‐grade units cropping out along the southern edge
of the Agly Massif by a very steep tectonic contact and from the southern Salvezines Mesozoic cover by a
complex and highly deformed tectonic zone (Demange & Pascal, 1979; Pascal, 1979) incorporating a
400‐m‐wide body of serpentinized mantle peridotite. (ii) The Bas‐Agly syncline to the ENE is in tectonic
contact with the Agly Massif on its southwesternmost limb. (iii) The Saint‐Paul‐de‐Fenouillet synform to
the north is overthrust by the Agly Massif on its southern limb and by the Bas‐Agly syncline in the
Tautavel region (Tautavel syncline).
3.3.2. Mesozoic Metamorphism and Metasomatism
The Bas‐Agly and Boucheville synforms, which lie to the north and south of the Agly‐Salvezines block
(Figure 2), are assigned to the Metamorphic Internal Zone as both record HT metamorphism (400 to
600 °C; Chelalou, 2015; Ducoux, 2017; Golberg & Leyreloup, 1990; Vauchez et al., 2013). The most recent
radiometric dating studies of this metamorphism in the Liassic‐Albian strata (Albarède & Michard‐Vitrac,
1978; Golberg & Maluski, 1988; Montigny et al., 1986) present ages of 95–90 Ma (late Cenomanian–late
Turonian), coeval with magmatic activity in the Corbières region, further northeast. On the basis of
RSCM paleotemperature data in the Boucheville synform, Chelalou et al. (2016) propose that temperature
increases downward through the complexly folded and poorly dated succession, giving an estimated vertical
geothermal gradient of 80 °C/km for the Cenomanian–Turonian peak HT‐LP event. In contrast, in the sim-
pler structure of the Bas‐Agly syncline, highest temperatures are recorded in the youngest strata (Berriasian
to Albian) with lower temperatures in older strata (350–500 °C; Chelalou, 2015; Ducoux, 2017; Golberg &
Leyreloup, 1990; Vauchez et al., 2013) and underlying pre‐Mesozoic rocks of the Agly Massif (maximum
temperatures of 350 °C in the Calce region, Figure 2; Clerc et al., 2015; Ducoux, 2017). Small pockets of lower
Liassic to lowermost Cretaceous marbles, scapolite‐bearing limestones, and breccias occur along fault zones
within the Agly Massif itself (Berger et al., 1993; Fonteilles et al., 1993), recording temperatures as high as
400–450 °C (Serre de Vergès, Figure 2; Ducoux, 2017). In contrast, further north, the Barremian–Albian
sediments of the Saint‐Paul‐de‐Fenouillet syncline record markedly lower temperature metamorphism
(Tectonics 10.1029/2018TC005254
low‐temperature metamorphism exists, so that it may be attributed either to synrift events or to
Pyrenean orogenesis.
Albitization is almost exclusively confined to the northern Agly‐Salvezines compartment (Boulvais et al.,
2007; Pascal, 1979). Upper crustal granites of this compartment show hydrothermal alteration to albitite
dated at 117–96 Ma (1.5–2 kbar and 350–450 °C; Boulvais et al., 2007; Demange & Pascal, 1979; Guinault
et al., 2016; Pascal, 1979; Poujol et al., 2010). This precedes the Cenomanian–Turonian peak HT‐LP event
in the surrounding Boucheville and Bas‐Agly synforms. The pressure conditions for albitization suggest a
crustal depth of 5 to 8 km. If one considers that temperature conditions correspond to background tempera-
tures, estimates ranging from 40 to 90 °C/km are derived for the geothermal gradient that prevailed over the
Agly‐Salvezines block during the Aptian to Cenomanian.
The reduced crustal profile of the Agly‐Salvezines block has been alternatively interpreted as due to (i) late
Variscan to post‐Variscan extension, following either orogenic crustal thickening (Bouhallier et al., 1991)
or transpressive thermal doming (Olivier et al., 2004, 2008), or (ii) Early Cretaceous extension based on
directional similarities of basement stretching to Early Cretaceous stretching of the Agly Mesozoic cover
(Delay, 1989; Delay & Paquet, 1989; Paquet & Delay, 1989; Paquet & Mansy, 1991, 1992). While not favor-
ing either interpretation, the work by R. Nicolas cited in Clerc et al. (2015) shows that partial (re)activation
of Variscan mylonitic extensional shear zones under greenschist facies conditions in the southern Agly
compartment (300–400 °C, 2–3 kbar) was synchronous with Early Cretaceous rifting. These pressure
conditions suggest a crustal depth of 8 to 12 km for the Agly southern compartment, which is consistent
with its expected structural position with respect to the northern granitic compartment at that time.
However, if one considers that the temperature conditions of shear zone recrystallization correspond to
background temperatures during the Early Cretaceous, this gives estimates of only 30–50 °C/km for the
geothermal gradient that prevailed over the Agly‐Salvezines block. These estimates are lower than but
overlapping with those obtained for albitization in the northern Agly‐Salvezines compartment over the
same period of time.
3.3.3. Thermal State of the Agly‐Salvezines Block at Onset of Pyrenean Convergence
Although the Agly Massif lies between two sectors of the Metamorphic Internal Zone, which both record
temperatures of up to 500–600 °C, the basement rocks of the massif did not record any post‐Variscan tem-
perature higher than the greenschist facies temperature conditions reached during the Variscan.
Moreover, the vertical geothermal gradient that prevailed specifically over the Agly‐Salvezines block during
the Early Cretaceous (rifting, albitization, and mylonitic reactivation) is poorly constrained, with overall
estimates ranging from 30 to 80 °C/km and varying between the northern and southern compartments.
Many processes have been proposed as the heat source for HT conditions recorded in Mesozoic strata: (i)
exhumed mantle (Clerc & Lagabrielle, 2014; Clerc et al., 2015, 2016; Chelalou et al., 2016; Vauchez et al.,
2013), and/or (ii) high thermal fluxes along deeply rooted strike‐slip faults (Clerc et al., 2015, 2016;
Vaudin, 1982), and/or (iii) heat transfer by fluid circulation (Boulvais, 2016; Clerc et al., 2015, 2016;
Dauteuil & Ricou, 1989). Along with evidence of extensional shear zones possibly active during the
Cretaceous on the northern side of the Agly Massif (Paquet & Mansy, 1991), the reduced crustal section of
the Agly Massif, and the presence of a block of mantle peridotite in the Boucheville synform, these features
have led recent workers to position the Agly‐Salvezines block at the distal edge of the Aptian–Cenomanian
rift system, where deep crustal levels and subcontinental mantle rocks were exhumed (Chelalou et al., 2016;
Clerc et al., 2016; Clerc & Lagabrielle, 2014; Ducoux, 2017; Vauchez et al., 2013). While all models of these
authors consider the Boucheville and Bas‐Agly synforms as autochtonous depocenters to the north and
south of the Agly Massif, they differ in that they propose either (i) unroofing of the Agly‐Salvezines basement
by Albian times (Clerc & Lagabrielle, 2014; Ravier, 1957; Vauchez et al., 2013) or (ii) burial of the massif
below a continuous and thick Apto‐Albian depocenter (Chelalou et al., 2016; Clerc et al., 2015, 2016;
Ducoux, 2017; Olivier, 2013). The former implies that the allochthonous Agly Mesozoic cover in the Calce
region has been almost fully preserved since the Albian. In contrast, the latter is supported by the absence
of any recorded clasts of Mesozoic or pre‐Mesozoic lithologies in the upper Aptian‐to‐lower Cenomanian
fine‐grained distal clastic marine sediments in either the Bas‐Agly or Saint‐Paul‐de‐Fenouillet depocenter
that could have been locally derived from the emerging massif (Berger et al., 1993). It also favors the
proposition of Clerc and Lagabrielle (2014) who argue that Albian sediments in Bas‐Agly and Boucheville
experienced HT conditions below a possibly 3‐ to 6‐km‐thick cover of Albo‐Cenomanian flysch. As
TERNOIS ET AL. 1314Tectonics 10.1029/2018TC005254
highlighted by Olivier (2013), currently available field and geochemical data do not allow a conclusive deci-
sion on this point.
New data are therefore required to constrain the structural and thermal evolution of the Agly‐Salvezines
block during early orogenesis and to define the thermal state of the crustal block before the onset of orogen-
esis. If the block was part of the distal margin of an Aptian–Cenomanian rift system, possibly hyperextended,
we would expect rift‐related heating and subsequent postrift cooling. However, we do not know the depth of
burial of the Agly‐Salvezines block during the Early Cretaceous, nor do we know the extent of the HT‐LP
metamorphic event affecting the block. With low‐temperature thermochronology we aim to distinguish
between burial below a thin sedimentary cover (relatively cold signature) or a thick cover (relatively hot sig-
nature). We also aim to get a better sense of background temperatures through time and investigate whether
the Early Cretaceous metamorphism was focused in fault zones and/or localized in certain areas. The data
presented in this study aim to answer some of these critical issues regarding the rift‐to‐collision transition.
4. (U‐Th)/He Thermochronology Analyses
4.1. Sampling Strategy
We collected a total of 17 samples from the Agly and Salvezines massifs: 10 granitic samples, mostly in the
northern, upper compartment, and 1 detrital and 6 gneissic samples in the southern, lower compartment
(Figure 2). Four granitic samples (15‐AG‐08, NP15‐AM‐A07, NP15‐AM‐A08, and NP15‐AM‐A10) were col-
lected in the Upper Carboniferous Saint‐Arnac granite pluton in order to document the top, center, and base
of the pluton (Olivier et al., 2004, 2008). Three samples with similar facies (G3, G4 and G5) were collected in
the Salvezines albitized granite pluton. Sample 15‐AG‐01 was collected from sillimanite‐ and muscovite‐
bearing Ordovician micaschists, and sample 15‐AG‐07 was collected in the Rivérole orthogneiss. Four
samples were collected in the Bélesta Unit: samples C13MA‐1B and 15‐AG‐06 are gneissic, and samples
NP15‐AM‐A11 and 15‐AG‐02 present granitic textures. Finally, three gneissic samples (C13MA‐2A,
15‐AG‐03, and 15‐AG‐04) were collected in the Caramany Unit, and sample 15‐AG‐05 was collected in the
Upper Carboniferous Ansignan charnockitic pluton.
4.2. Zircon (U‐Th)/He Methodology
Zircon grains from the 50‐ to 250‐μm fraction were isolated for (U‐Th)/He dating analyses using conven-
tional mineral separation techniques, including crushing, grinding, sieving, and water table concentration,
standard heavy‐liquid density, and magnetic susceptibility separation techniques. Zircon grains were subse-
quently handpicked for analysis according to the following conditions: (i) of bipyramidal shape (euhedral),
unbroken and visibly fracture free, (ii) of any size (length, width, and thickness) >60 μm, (iii) of any
transparency/opacity, and (iv) inclusion free when inspected under a binocular microscope.
The (U‐Th)/He dating method is based on the radioactive decay of 238U, 235U, and 232Th (decay series) into
4
He or alpha particles. Conventional bulk single‐grain ZHe analyses were performed conjointly at the Centre
de Recherches Pétrographiques et Géochimiques (CRPG; Nancy, France) and at the UTChron facilities at
the Jackson School of Geosciences (University of Texas, Austin, USA). A bulk (U‐Th)/He age is derived from
measurements of parent and daughter nuclide abundances in a single, entire crystal (aliquot). It must be
acknowledged that as is typical for most, if not all, conventional bulk ZHe dating studies, the degree of
zonation within each of our dated grains was not measured due to intrinsic analytical protocol
limitations. This will be discussed in section 5. Nine samples were analyzed at the CRPG following proce-
dures described in Pik et al. (2003) and Godard et al. (2009) for 4He extraction and Tibari et al. (2016) for bulk
parent U, Th, and Sm concentration measurements. The other eight samples were analyzed at the Jackson
School of Geosciences following the procedures described in Wolfe and Stockli (2010). Based on replicate
analysis of the Fish Canyon Tuff ZHe age standard, analytical uncertainties arising from the two protocols
for uncorrected ZHe ages are estimated at 6% (1σ) and 4% (1σ), respectively.
Given the traveling distances of 4He or alpha particles produced by radioactive decay of 238U, 235U, and 232Th
(~16.7, ~19.6, and ~19.3 μm, respectively; Farley et al., 1996), these particles can be ejected from the outer
~20 μm of the grain. As a result, ZHe ages were morphometrically corrected for alpha‐ejection (the ejection
correction parameter FT) according to Ketcham et al. (2011). We considered our zircon grains as equidimen-
sional bipyramidal solids, with width, a, equal to thickness, b, and the number of pyramids, Np, equal to 2.
TERNOIS ET AL. 1315Tectonics 10.1029/2018TC005254
Polynomial fit parameters for surface‐area‐to‐volume ratio, β, were those of Hourigan et al. (2005) for prisms
with pyramids. As the Agly‐Salvezines rocks are known to have experienced Carboniferous HT‐LP meta-
morphism, we expected ZHe ages younger than 500 Ma. Therefore, we considered that the contribution of
235
U to the total 4He concentration measured was negligible, and we decided to keep using FT as the linear
correction to the measured age as expressed by Farley et al. (1996).
4.3. Apatite (U‐Th)/He Methodology
After mineral separation, apatite grains were handpicked for analysis according to the following conditions:
(i) hexagonal shape (euhedral), unbroken and visibly fracture free, (ii) of sufficient size (width and thick-
ness) >70 μm, and (iii) absolutely inclusion free when carefully inspected under a binocular microscope
at ×180 immersed in alcohol. Selection of apatite absolutely free of U‐ and Th‐bearing inclusions represents
the biggest hurdle to accurate and reliable AHe age determinations. All conventional single‐grain AHe ana-
lyses were performed at the UTChron facilities at the University of Texas, following procedures described in
Prior et al. (2016).
Similar to ZHe ages, AHe ages were corrected morphometrically for alpha‐ejection (FT correction) according
to Ketcham et al. (2011) due to long attenuation distances of radiogenic 4He or alpha particles produced from
238
U, 235U, and 232Th (~19.7, ~22.8, and ~22.5 μm, respectively; Farley et al., 1996). After in vacuo total‐
fusion degassing by UV‐laser heating for 5 min at 1070 °C and cryogenic purification, 4He concentrations
of single‐grain apatite were determined using quadrupole mass spectrometry by isotope dilution (3He spike).
Degassed apatite aliquots are retrieved, dissolved in 30% nitric acid, spiked with an isotopically enriched tra-
cer, and analyzed for U, Th, and Sm concentrations using a Thermo Element2 inductively coupled plasma
mass spectrometer fitted with a microcentric nebulizer. Based on replicate analysis of the Fish Canyon
Tuff AHe age and other fast‐cooled in‐house volcanic apatite standards, analytical uncertainties for uncor-
rected AHe ages are estimated at 3% (1σ).
5. (U‐Th)/He Results
Our 17 samples provided a total of 94 zircon crystals satisfying the suitability tests for further investiga-
tion. Three of these samples also provided two suitable apatite crystals each. Our (U‐Th)/He data are
reported in Data Set S1 in the supporting information. Zircon grains yield FT‐corrected He ages ranging
from 39.0 ± 2.3 to 117.1 ± 7.0 Ma with intrasample grain age variation from ∼19% to ∼93%. Our AHe
ages range from 40.1 ± 2.4 to 55.0 ± 3.3 Ma. ZHe ages older than the Paleocene account for ∼87% of
the Agly‐Salvezines data set. Ten samples yield middle Aptian to late Turonian (117–90 Ma) ZHe ages,
hereafter referred to as middle Cretaceous ages. These ages account for ∼22% of the data set. Only 2
zircons (∼2% of the data set) yield Eocene ages, which fall into the range of our AHe ages and the
AFT ages of Yelland (1991) and Gunnell et al. (2009) for the Agly Massif (40.3 ± 2.6 to 49.0 ± 1.8 Ma
for mean track lengths of 13.19–14.18 μm).
Our ZHe and AHe data set shows a high degree of age variability due to significant intrasample variation,
which is larger than would be expected given the analytical precision. It has been demonstrated that this
could result from the inability of a single set of kinematic parameters for He diffusion to describe the natural
thermochronometric behavior of zircon (Cherniak et al., 2009; Guenthner et al., 2013; Ketcham et al., 2013;
Reiners et al., 2002, 2004) and apatite (Flowers et al., 2007; Gautheron et al., 2009; Shuster et al., 2006).
Grain‐specific characteristics such as grain size, parent nuclide zoning, and radiation damage are among
the most important known factors that influence helium diffusivity, all the more exacerbated in zircon
(Dobson et al., 2008; Guenthner et al., 2013; Hourigan et al., 2005; Reiners et al., 2004; Wolfe & Stockli,
2010). It is therefore not unreasonable to expect outliers for crystals with very low helium retentivity. As
the two Eocene ZHe ages fall into the range of both previously published AFT data and our AHe data, they
are considered as outlier data points and will subsequently be ignored.
Systematic correlations between ZHe ages and equivalent spherical radius (an approximation of grain size)
are not apparent in any of our individual sample zircon data sets (Figure S1 in the supporting information).
Although no clear relationships between these two variables are found for zircon, we will nevertheless con-
sider the grain size effect in later discussions. Regarding our AHe ages, it is not reasonable to search for any
correlation with only two ages per sample, and we will therefore focus mainly on zircon ages. Given that
TERNOIS ET AL. 1316Tectonics 10.1029/2018TC005254
zonation of U and Th is a potential source of age variation that has its strongest effect for grains with an
∼60 μm equivalent spherical radius when high eU cores or high eU rims occupy the inner or outer third
of the grain, respectively (the zircon impact index/factor/number; Guenthner et al., 2013, 2015), our lack
of observations of zircon zonation for grains with a similar equivalent spherical radius on average
(∼65 μm) may appear to be an issue. In particular, Tournaire Guille et al. (2018) have recently shown that
typical zircons collected from some of our sample locations show pre‐Variscan‐inherited cores and Upper
Carboniferous metamorphic overgrowths that may occupy one third of the grains. However, Odlum and
Stockli (2017), who analyzed zircons from eight of the samples presented in our study for U/Pb dating (solid
triangles in Figures 2 and 3), show that zircon zonation is only observed in gneissic samples 15‐AG‐06 and
15‐AG‐03. In addition, the U and Th abundance data for zoned grains presented in the two previously
mentioned studies give a good sense of the style of zonation in all the grains analyzed. First, very few meta-
morphic overgrowths have been documented in typical zircons collected from the igneous sample locations.
Second, these data indicate rim eU enrichment of less than an order of magnitude on average for zoned
grains in the Bélesta and Caramany units, therefore likely placing the level of zonation for our equivalent
grains between the “mildly problematic and worst case” scenarios of Guenthner et al. (2015; factor of enrich-
ment of 2 and 10, respectively, for zircon grains systematically zoned in the same fashion with high eU rims
occupying the outer third of the grain). These observations suggest that heterogeneous distribution of U and
Th in our zircons, albeit likely contributing to some of the observed age irreproducibility in each sample,
cannot alone explain the large spread in our ZHe ages. The remainder of this results section, and more
generally this study, will therefore be mainly focused on the third main source of age variation in our
ZHe data set: radiation damage (Guenthner et al., 2013, 2014, 2015, 2017).
Single‐grain He ages are plotted against the bulk effective uranium concentration, eU, in Figure 3a (eU
[ppm] = U [ppm] + 0.235 * Th [ppm]). Bulk eU is calculated for each grain, providing a proxy for both
the production rate of alpha particles from decay of U, Th, and their daughter products (Sm being a minor
contributor to the alpha particle production budget) and the relative amount of radiation damage accumu-
lated in each crystal of a sample that has not experienced annealing (for apatite: Flowers et al., 2007; Shuster
et al., 2006; for zircon: Guenthner et al., 2013).
Our AHe ages are older at higher bulk eU. The age variability in our AHe data set is similar to that previously
reported in other geological settings (Gautheron et al., 2009, and references therein). In particular, AHe ages
at high eU concentrations may be found to be as old as or even a little older than AFT ages for the same geo-
logical province. It has been demonstrated that alpha‐recoil damage induced by recoil of U and Th decay
series during emission of alpha particles and eU concentration govern He diffusion within apatite crystals,
and thus the resulting AHe ages (Flowers et al., 2007; Gautheron et al., 2009; Shuster et al., 2006). The
amount of alpha‐recoil damage increases with eU, enhancing He retention for apatite crystals with higher
eU and extending crystal temperature sensitivity from ∼40–80 °C (Crowley et al., 2002; Farley, 2000;
Farley et al., 1996; Stockli et al., 2000) up to (sometimes even higher than) that of the AFT system
(∼100 °C; Gallagher et al., 1998).
Similar age variability to that in our ZHe data set is reported in other data sets from the Pyrenees and from
other similar orogenic settings (e.g., Guenthner et al., 2014, 2015, 2017; Vacherat et al., 2014, 2016). It has
been demonstrated that this results from the inability of a single set of kinematic parameters for He diffusion
to describe the natural thermochronometric behavior of zircon (Cherniak et al., 2009; Guenthner et al., 2013;
Ketcham et al., 2013; Reiners et al., 2002, 2004). In particular, Guenthner et al. (2013) have recently proposed
that such age variation in data sets illustrates the damage‐dependent sensitivity of He in zircon to thermal
processes. Similarly to the AHe system, these authors have established a zircon radiation‐damage and
annealing model that quantitatively describes the effects of radiation damage on He diffusion in zircon as
a function of time and temperature. They show the clear dependency of He diffusivity on radiation damage
by parameterizing the relationship between damage and diffusivity. Their model considers a partial anneal-
ing zone of 310–223 °C and the 8‐orders‐of‐magnitude variation in diffusivity across the radiation dose
spectrum. It also integrates the damage‐based diffusion kinetics to calculate ZHe closure temperature as a
function of the alpha dose. The model thus predicts that the ZHe system in damaged zircons (damage
amounts ranging from 1 × 1016 to 1 × 1018 α/g) is sensitive to temperatures of 140 to 200–250 °C. This is
in good agreement with previously documented temperature ranges for the ZHe partial retention zone
(Reiners et al., 2002; Stockli, 2005; Tagami et al., 2003; Wolfe & Stockli, 2010).
TERNOIS ET AL. 1317Tectonics 10.1029/2018TC005254
Figure 3. (a) Zircon (left and middle inserts) and apatite (right insert) He age‐eU correlations for the Agly‐Salvezines data set, divided with the sample and group
color scheme used in Figure 2. ZHe age data for Gp1 and Gp2 are plotted separately. Gray bars indicate the ages of major thermal and structural events from
literature. (b) Density graphs of the single‐grain (U‐Th)/He ages obtained in this study (2‐D Kernel density estimation via the Parzen‐Rosenblatt window
method). 1: Albarède and Michard‐Vitrac (1978); 2: Montigny et al. (1986); 3: Golberg and Maluski (1988); 4: Yelland (1991); 5: work by R. Nicolas cited in Clerc
et al. (2015); 6: Maurel (2003); 7: Boulvais et al. (2007); 8: Denèle (2007); 9: Maurel et al. (2008); 10: Gunnell et al. (2009); 11: Poujol et al. (2010); 12: Guinault
et al. (2016).
In the case of ZHe thermochronology, age‐eU plots have recently proved to be very useful to gain insight into
the thermal history of a sample or a group of samples in several studies. It has been demonstrated that a
sample or a group of samples may yield specific age‐eU correlations, positive or negative, even both, as a
result of its entire thermal history (e.g., Anderson et al., 2017; Guenthner et al., 2013, 2014, 2015, 2017;
Hart et al., 2017; Johnson et al., 2017; Powell et al., 2016; Thomson et al., 2017; Vacherat et al., 2014,
2016). In our He age‐eU plots (Figure 3) we split the Agly‐Salvezines sample collection into two distinct
sample groups (Gp1 and Gp2) based on the sample rock type (igneous or metamorphic) and sample location
(or inferred pre‐Early Cretaceous sample depth: northern, upper compartment versus southern, lower
compartment), irrespective of either the intrasample variation criterion or grain characteristics. This aimed
at separating rocks that had, as a first approximation, distinct pre‐Early Cretaceous history and therefore
crystals with distinct grain‐specific characteristics. Considerations of eU and sample groups are discussed
TERNOIS ET AL. 1318Tectonics 10.1029/2018TC005254
in more detail in later sections. Error bars shown here represent the 1σ analytical uncertainty for each single‐
grain aliquot. Results for the two massifs, Agly and Salvezines, are presented together.
The first group (Gp1) comprises the seven igneous samples located in the northern compartment of Agly‐
Salvezines (Figure 2). It is characterized by ZHe ages ranging from 56.6 ± 3.4 to 110.2 ± 6.6 Ma.
Intrasample age variability is the same as that of the whole ZHe data set. ZHe ages are associated with bulk
eU concentrations ranging from 183 to 1,633 ppm. Individual age and eU ranges are similar in the error bars
from one sample to another, with the exception of sample NP15‐AM‐A10, which yields the oldest (Apto‐
Albian) ZHe ages for this group. Four samples (NP15‐AM‐A10, 15‐AG‐08, G3, and G5) have a clear negative
age‐eU correlation as ages decrease from ∼90 ± 5 Ma to ∼65 ± 5 with increasing eU (Figure S2). Sample
NP15‐AM‐A07 may have a positive age‐eU correlation, as ages increase from ∼75 ± 5 Ma to ∼95 ± 5 with
increasing eU.
The second group (Gp2) comprises the 10 samples from the southern compartment of Agly (Figure 2). Both
the intrasample grain variations and the ZHe age range of these samples are similar to those of Gp1 (age
variability of ∼19% to ∼71%; ages of 63.8 ± 3.8 to 117.1 ± 7.0 Ma). Nonetheless, the ZHe age range is mostly
confined to relatively low bulk eU concentrations (152–879 ppm) and a larger proportion of ages clustering
around 60–75 Ma is found compared with Gp1. Five samples (15‐AG‐02, 15‐AG‐03, 15‐AG‐04, 15‐AG‐05,
and 15‐AG‐06) have a positive age‐eU correlation as ages increase from ∼70 ± 5 to ∼100 ± 5 Ma with increas-
ing eU (Figure S2). Three samples (NP15‐AM‐A11, C13‐MA‐2A, and 15‐AG‐07) may have a negative age‐
eU correlation.
Second‐order scatter along the dominant age‐eU trends is observed and likely results from the combined
effects of grain size and, more importantly, intragranular U and Th heterogeneities within each grain.
Nonetheless, it is clear that samples from the same group show similar individual zircon eU ranges and
age‐eU correlations (first‐order negative correlations for Gp1 samples and positive correlations for Gp2
samples). We therefore argue that all samples of Gp1 experienced the same thermal history since the ZHe
clock started. The same can be concluded for all samples of Gp2.
For each group, we combined individual sample data sets into a group data set and generated a two‐
dimension (age and eU) kernel density estimation of the data (Figure 3b). The statistical distribution of
Gp1 ages (37 ZHe ages) highlights two significant age clusters. The main age cluster (80 Ma) is found at
eU concentrations of ∼400 ppm. The second age cluster (70 Ma) corresponds to zircon crystals with eU con-
centrations ranging from ∼1,000 to ∼1,600 ppm. An additional age cluster, although poorly defined, may be
centered on 110 Ma and 300–400 ppm. The age distribution in Gp2 (55 ZHe ages) suggests two age clusters at
low eU concentrations (1,000 ppm). Both sample groups yield overlapping single‐grain ZHe ages over the whole spread in bulk
eU concentrations. Although 82–95% of ages from the Gp1 data set are younger than 90–95 Ma (compared
with 64–79% for Gp2), arguing for distinct maximum ZHe ages of Gp1 and Gp2 is hazardous. Age distribu-
tions (Figure 3b) do show an age cluster centered on 117–90 Ma (eU of ∼300–500 ppm) for both groups.
The situation is the same for most of ZHe ages around 65–75 Ma at low eU (∼300–400 ppm) for Gp2 com-
pared with Gp1 and around 70 Ma at high eU (1,000–1,500 ppm) for Gp1 compared with Gp2. These clusters
are the cornerstones of the overall ZHe age‐eU correlations observed for each group. First, these slight pro-
portional differences could simply be the result of undersampling. Second, Odlum and Stockli (2017) and
Tournaire Guille et al. (2018) have shown that incorporation of U and Th into zircon grains of the Agly
Massif depends on the complexity of the grain crystallization history. In particular, typical zircons collected
from the southern sample locations show systematic rim eU enrichment, so that they may have distinct
domains with variable radiation damage, and therefore variable diffusivities. Consequently, this not only
suggests that the ages these grains yield will be more dispersed or randomly distributed than ages of unzoned
grains (igneous, northern compartment), but it also implies that they will probably plot at lower eU than
their counterparts. One has then statistically more chance of handpicking zircon crystals with low bulk
eU concentrations (or less chance in case of high bulk eU) in rocks from the southern compartment than
in plutonic rocks from the northern compartment. Taken together, these two arguments could explain the
TERNOIS ET AL. 1319Tectonics 10.1029/2018TC005254
better documentation of positive age‐eU correlations for Gp2 and the poorer documentation of negative cor-
relations in Gp1. On another note, based on the Guenthner et al. (2013) model, one can expect older crystals
with longer exhumation histories to accumulate more radiation damage in the absence of annealing, shifting
maximum retentivity for He and associated older ZHe ages at lower eU concentrations. All our samples yield
older ages at similar bulk eU concentrations, most likely indicating homogeneous annealing (310–223 °C
and above) of radiation damage of zircons at some point in their thermal history (the Variscan orogeny
and the Aptian–middle Cenomanian HT‐LP event are the best candidates; see below). We therefore argue
that the reverse trends of age‐eU correlations for Gp1 and Gp2 do not necessarily arise from distinct thermal
histories for the two groups but may be simply due to some kind of sampling bias, as described above. The age
clusters at low eU (∼300–400 ppm) around 70 Ma for Gp2 and 80 Ma for Gp1 are most likely equivalent, pos-
sibly suggesting that rocks lying at different depths reached a particular closure temperature depth with
some 10 myrs of delay. In a wider perspective, our AHe ages are very close to each other and display as a
whole a positive age‐eU correlation, consistent with apatite radiation‐damage and annealing models. On
the basis of all the low‐temperature thermochronology similarities presented by the two groups, we expect
all samples from this study to have experienced the same thermal history, at least since the
middle Cretaceous.
6. Numerical (U‐Th)/He Age Modeling
Recent studies (e.g., Anderson et al., 2017; Guenthner et al., 2014, 2015, 2017; Johnson et al., 2017; Powell
et al., 2016; Vacherat et al., 2014, 2016) have shown that numerical modeling, incorporating apatite and
zircon radiation‐damage and annealing models for both the fission track and (U‐Th)/He systems, is
extremely valuable for understanding and interpreting data. It allows exploration of the response of different
thermochronometers to various thermal histories and ultimately provides more quantitative constraints on
the thermal history of a sample. To constrain the range of plausible t‐T paths producing either model age‐eU
correlations or single‐grain He ages that most closely match the data presented in this study, we performed
both inverse modeling and forward modeling using the software program HeFTy (Ketcham, 2005).
6.1. Data Inverse Modeling Approach Using HeFTy
The new ZHe and AHe data presented in this study complete, and agree with, the sparse AFT data obtained
by Yelland (1991) and Gunnell et al. (2009) for the Agly‐Salvezines block. Altogether, these data document
the time‐temperature history of the crustal block in the overall temperature range of 45–250 °C (lower bound
of the AHe partial retention zone‐upper bound of the ZHe partial retention zone; e.g., Guenthner et al., 2013;
Stockli et al., 2000).
6.1.1. HeFTy Thermochronology Model Parameterization
To explore plausible t‐T spaces consistent with our data, we first used the inverse modeling approach. The
number of thermochronometric models that can be input into HeFTy for inverse modeling is relatively lim-
ited. In particular, as already mentioned in Vacherat et al. (2014) and Powell et al. (2016), the full range of
single‐grain eU‐dependent He ages cannot be used, so that representative individual ZHe and AHe ages
must be chosen. All the model inputs and constraints used for our inverse modeling simulations are reported
in Table S1, following the model input reporting protocol of Flowers et al. (2015).
Three ZHe age‐eU couples were used as parameterizations of three HeFTy ZHe models, each of them
representing one of the three age clusters shared by the two groups. The first cluster is defined around
75 Ma (± 9.4 Ma, 12%, 2σ analytical uncertainty) at low eU concentrations (300–500 ppm; HeFTy ZHe
model 1). The second cluster concerns ZHe ages around 70 Ma (± 12%, 2σ analytical uncertainty) at
relatively high eU concentrations (1,500 ppm; HeFTy ZHe model 2). The last cluster represents middle
Cretaceous ages (105 Ma, ±12%, 2σ analytical uncertainty, ∼400 ppm; HeFTy ZHe model 3). A similar
reasoning was applied for AHe data with two AHe models characterizing ages either at low or at high
eU concentrations (HeFTy AHe models 1 and 2). Finally, two AFT models were considered, based on
Gunnell et al. (2009; HeFTy AFT models AGLY1 and AGLY5).
To reproduce the range of our ZHe ages (∼70–110 Ma), our AHe ages (∼43–50 Ma), and published AFT ages
(∼40–50 Ma), the post‐Barremian time‐temperature history was first left free within a t‐T space limited by
the upper bound of the partial annealing zone of the Guenthner et al. (2013) model (∼300 °C) and surface
TERNOIS ET AL. 1320You can also read
